Compression: From Cochlea to Cochlear Implants (Springer Handbook of Auditory Research)

Citation preview

Springer Handbook of Auditory Research Series Editors: Richard R. Fay and Arthur N. Popper

Springer New York Berlin Heidelberg Hong Kong London Milan Paris Tokyo

This page intentionally left blank

Sid P. Bacon Richard R. Fay Arthur N. Popper Editors

Compression: From Cochlea to Cochlear Implants

With 62 Illustrations

13

Sid P. Bacon Department of Speech and Hearing Science and Psychoacoustics Laboratory Arizona State University Tempe, AZ 85287-1908, USA [email protected]

Richard R. Fay Department of Psychology and Parmly Hearing Institute Loyola University of Chicago Chicago, IL 60626, USA [email protected]

Arthur N. Popper Department of Biology and Neuroscience and Cognitive Science Program and Center for Comparative and Evolutionary Biology of Hearing University of Maryland College Park, MD 20742–4415, USA [email protected] Cover illustration: Functional anatomy of the mammalian cochlea. This figure appears on p. 20 of the text.

Library of Congress Cataloging-in-Publication Data Compression : from cochlea to cochlear implants / editor, Sid P. Bacon, Richard R. Fay, Arthur N. Popper. p. cm.—(Springer handbook of auditory research ; v. 17) Includes bibliographical references and index. ISBN 0-387-00496-3 (alk. paper) 1. Cochlea. 2. Compression (Audiology). 3. Auditory perception. 4. Cochlear implants. I. Bacon, Sid P. II. Fay, Richard R. III. Popper, Arthur N. IV. Series. QP471.2.C66 2003 612.8¢58—dc21 2003042430 ISBN 0-387-00496-3

Printed on acid-free paper.

© 2004 Springer-Verlag New York, Inc. All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer-Verlag New York, Inc., 175 Fifth Avenue, New York, NY 10010, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. 9 8 7 6 5 4 3 2 1

SPIN 10913048

www.springer-ny.com Springer-Verlag New York Berlin Heidelberg A member of BertelsmannSpringer Science+Business Media GmbH

This book is dedicated to Professor Neal F. Viemeister, whose numerous scientific contributions have led to a deeper understanding of auditory perception and its intricate relation to physiological processing at the auditory periphery.

This page intentionally left blank

Series Preface

The Springer Handbook of Auditory Research presents a series of comprehensive and synthetic reviews of the fundamental topics in modern auditory research. The volumes are aimed at all individuals with interests in hearing research including advanced graduate students, post-doctoral researchers, and clinical investigators. The volumes are intended to introduce new investigators to important aspects of hearing science and to help established investigators to better understand the fundamental theories and data in fields of hearing that they may not normally follow closely. Each volume is intended to present a particular topic comprehensively, and each chapter will serve as a synthetic overview and guide to the literature. As such the chapters present neither exhaustive data reviews nor original research that has not yet appeared in peer-reviewed journals. The volumes focus on topics that have developed a solid data and conceptual foundation rather than on those for which a literature is only beginning to develop. New research areas will be covered on a timely basis in the series as they begin to mature. Each volume in the series consists of five to eight substantial chapters on a particular topic. In some cases, the topics will be ones of traditional interest for which there is a substantial body of data and theory, such as auditory neuroanatomy (Vol. 1) and neurophysiology (Vol. 2). Other volumes in the series will deal with topics that have begun to mature more recently, such as development, plasticity, and computational models of neural processing. In many cases, the series editors will be joined by a co-editor having special expertise in the topic of the volume. Richard R. Fay, Chicago, Illinois Arthur N. Popper, College Park, Maryland

vii

This page intentionally left blank

Preface

One truly remarkable aspect of the auditory system is its ability to process sounds over an incredibly wide range of levels, on the order of 120 dB. For example, the loudness of a sound will increase with increasing level over this entire range. That the auditory system has an extraordinarily large dynamic range and is acutely sensitive to small changes in level throughout that range is an especially important aspect of hearing because it enables humans to process speech over a wide range of levels and to extract important information from the small changes in level that occur over that rather wide range. An interesting and important problem for hearing scientists has been to understand how the auditory system is able to process sound over such a wide dynamic range. As discussed in this volume, it is now believed that the enormous psychophysical dynamic range is accomplished via a form of compression that exists at the level of the cochlea. This peripheral auditory compression is the focus of this book. In the first chapter, Bacon provides a brief overview of peripheral compression, the perceptual consequences of this compression, the perceptual consequences of reduced or absent compression after cochlear hearing loss, and the signal-processing strategies used to compensate for this reduced compression. This chapter sets the stage for the remaining chapters, where those topics are covered in detail. In Chapter 2, Cooper describes peripheral auditory compression as observed in the mechanical responses of the basilar membrane as well as in the electrical responses of the auditory nerve. In Chapter 3, Oxenham and Bacon describe a wide range of perceptual consequences that are thought to result from peripheral auditory compression. Then in Chapter 4, Bacon and Oxenham discuss how perceptual processing can be affected by a reduction in or loss of compression after cochlear damage. In Chapter 5, Levitt discusses a form of signal processing (known as compression amplification) that is often used in hearing aids as a means to at least partially compensate for reduced peripheral compression in individuals with cochlear hearing loss. Finally, in Chapter 6, Zeng describes psychophysical research on individuals fitted with cochlear implants and discusses the theix

x

Preface

oretical implications of that research for peripheral compression and the practical implications for signal processing in cochlear implants. This volume can serve as a companion to several others in the Springer Handbook of Auditory Research. Many aspects of psychophysics were reviewed in Human Psychophysics (Vol. 3). A number of other chapters in the series are related to similar topics in other animal species. For example, psychophysics of fish and amphibians is considered in a chapter by Fay and Megela Simmons in Comparative Hearing: Fish and Amphibians (Vol. 11), on birds and reptiles in a chapter by Dooling et al. in Comparative Hearing: Birds and Reptiles (Vol. 13), and on dolphins and whales in a chapter by Nachtigall et al. in Hearing by Whales and Dolphins (Vol. 12). Sid P. Bacon, Tempe, Arizona Richard R. Fay, Chicago, Illinois Arthur N. Popper, College Park, Maryland

Contents

Series Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii ix xiii

Chapter 1 Overview of Auditory Compression . . . . . . . . . . . . . . . . Sid P. Bacon

1

Chapter 2 Compression in the Peripheral Auditory System . . . . . . Nigel P. Cooper

18

Chapter 3 Psychophysical Manifestations of Compression: Normal-Hearing Listeners . . . . . . . . . . . . . . . . . . . . . . . Andrew J. Oxenham and Sid P. Bacon

62

Chapter 4 Psychophysical Manifestations of Compression: Hearing-Impaired Listeners . . . . . . . . . . . . . . . . . . . . . . Sid P. Bacon and Andrew J. Oxenham

107

Chapter 5 Compression Amplification . . . . . . . . . . . . . . . . . . . . . . 153 Harry Levitt Chapter 6 Compression and Cochlear Implants . . . . . . . . . . . . . . . 184 Fan-Gang Zeng Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

221

xi

This page intentionally left blank

Contributors

Sid P. Bacon Psychoacoustics Laboratory, Department of Speech and Hearing Science, Arizona State University, Tempe, AZ 85287-1908, USA [email protected]

Nigel P. Cooper MacKay Institute of Communication and Neuroscience, Keele University, Keele, Staffordshire, ST5 5BG, UK [email protected]

Harry Levitt 998 SeaEagle Loop, Bodega Bay, CA 94923-0610, USA [email protected]

Andrew J. Oxenham Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA [email protected]

Fan-Gang Zeng University of California, 364 Med Surge II, Irvine, CA 92697, USA [email protected]

xiii

This page intentionally left blank

1 Overview of Auditory Compression Sid P. Bacon

1. Introduction One truly remarkable aspect of the auditory system is its ability to process sounds over an incredibly wide range of levels. This range for humans is illustrated in Figure 1.1. The bottom curve shows the sound pressure levels at which pure tones of a wide range of frequencies can just be detected. In other words, it shows the absolute thresholds for human hearing. This curve is sometimes referred to as an audibility curve. Everything below the curve is inaudible, whereas everything above it is audible.1 The top curve shows the sound pressure levels at which the pure tones are uncomfortably loud. The area between these two curves represents the so-called dynamic range of hearing. Although the dynamic range varies with frequency, the dynamic range for a tone with a frequency of around 1,000 Hz is at least 120 dB. This corresponds to a truly impressive dynamic range of 1012 intensity units (W/m2). The loudness of a sound will increase with increasing stimulus level throughout the range, and it may continue to increase at even higher levels, although few are willing to find out! Not only does the sensation of loudness grow with increasing level over this range but humans can also detect small (approximately 1 dB) changes in level over this entire range (Viemeister and Bacon 1988). Thus the auditory system has an extraordinarily large dynamic range and is acutely sensitive to small changes in level throughout that range. This is an especially important aspect of human hearing. It enables humans, for example, to accommodate the wide range of levels in speech that, when one considers both the changes in average level and the changes in instantaneous level, is on the order of 60–80 dB for everyday listening situations (e.g., Levitt 1982). Furthermore, it enables humans to extract important information from the small changes in level that occur over that wide range. 1

This is, however, not strictly true. The absolute threshold is defined statistically as the sound pressure level necessary to detect the sound a certain percentage of times (e.g., 70%). 1

2

S.P. Bacon

Figure 1.1. Bottom solid line: levels at which pure tones of various frequencies can be just detected; top solid line: levels at which the tones become uncomfortably loud. The area in between the 2 curves represents the dynamic range of hearing. SPL, sound pressure level.

An interesting problem for hearing scientists has been to understand how the auditory system is able to process or encode sound over such a large range of levels. The problem is particularly apparent when one considers that the typical dynamic range of auditory nerve fibers is on the order of 20–40 dB (for reviews, see Ruggero 1992; Cooper, Chapter 2). This substantial discrepancy between the behavioral or psychophysical dynamic range and the physiological dynamic range has been referred to as the “dynamic range problem” (Evans 1981). It has been the focus of both psychophysical and physiological research (for reviews, see Viemeister 1988a,b; Delgutte 1996). As discussed below (Section 2), it is now believed that the enormous psychophysical dynamic range is accomplished via a form of compression that exists at the level of the cochlea. This peripheral auditory compression is the focus of this book. The various chapters describe not only the cochlear processing underlying compression (Cooper, Chapter 2) but also the wide range of perceptual consequences that are thought to result from this compression (Oxenham and Bacon, Chapter 3) as well as the perceptual consequences of reduced or absent compression (Bacon and Oxenham, Chapter 4). This book also provides a discussion of the rehabilitative or signal-processing strategies used to compensate for the lack of peripheral auditory compression in individuals fitted with either hearing aids (Levitt, Chapter 5) or cochlear implants (Zeng, Chapter 6). The

1. Overview of Auditory Compression

3

purpose of this chapter is to provide a brief overview of these topics and hence set the stage for the more detailed discussions that follow.

2. Cochlear Compression The mammalian cochlea consists of two distinct types of sensory cells that run almost its entire length (for a review of cochlear anatomy, see Slepecky 1996). There is one row of inner hair cells and generally three rows of outer hair cells. The hair cells and their supporting cells lie directly above the basilar membrane. The vast majority (90–95%) of the auditory nerve fibers synapse directly with the inner hair cells, whereas the remainder synapse with the outer hair cells. It is generally believed that all recordings of neural activity from the auditory nerve fibers reflect the responses of fibers innervating inner hair cells (e.g., Liberman 1982; Liberman and Oliver 1984). Furthermore, those fibers that synapse with outer hair cells are apparently unresponsive to sound (Robertson 1984). Thus the neural code at the auditory nerve directly reflects the response properties solely of the inner hair cells. One might ask, then: What is the role of the considerably more numerous outer hair cells in hearing? Over the past 20 or so years, tremendous insight into this question has been gained (for reviews of cochlear mechanics and hair cell physiology, see Dallos 1988, 1992; Kros 1996; Patuzzi 1996; Robles and Ruggero 2001; Cooper, Chapter 2). The results from a wide range of scientific approaches clearly indicate that the outer hair cells are crucial for normal auditory function. Indeed, it is now understood that these cells are responsible for our exquisite sensitivity and superb frequencyresolving capabilities. Furthermore, the importance of these cells is underscored by the often severe auditory-processing difficulties experienced by individuals with outer hair cell damage (for a review of the perceptual consequences of hair cell damage, see Moore 1998). The outer hair cells are also responsible for various nonlinear phenomena that can be observed, for instance, in the mechanical response of the basilar membrane and the neural response of the auditory nerve. Moreover, psychophysical correlates of these nonlinear phenomena can be measured in human listeners. These nonlinear phenomena include the generation of distortion products (summation and difference tones) and two-tone suppression (see Cooper, Chapter 2) as well as a compressive growth of response, which is the focus of this book. The remainder of this section considers this compressive growth as it is observed at the basilar membrane. The basilar membrane vibrates in a characteristic manner in response to sound (for a review of cochlear mechanics, see Patuzzi 1996). In particular, there is a wave of displacement that travels from the base of the cochlea to the apex. It increases in magnitude until it reaches a peak at some place on the basilar membrane, and then it dies off rather abruptly. This is known as

4

S.P. Bacon

the traveling wave. Much of what is known about the traveling wave comes from the pioneering work on cochlear mechanics by Georg von Békésy, for which he won the Nobel Prize in Physiology or Medicine in 1961 (for a summary of that work, see von Békésy 1960). The place where the traveling wave reaches its peak depends on the frequency of the sound. High frequencies peak near the base of the cochlea, whereas low frequencies peak near the apex. Indeed, there is a frequency-to-place mapping along the basilar membrane that is known as tonotopic organization. This organization is maintained throughout the auditory system. Thus each place along the basilar membrane responds best to only one frequency, although it will respond to other frequencies as well. This frequency is usually referred to as the characteristic frequency (CF).2 To gain an appreciation of the compression that exists at the basilar membrane, it is necessary to consider how the magnitude of basilar membrane vibration at a given point along the membrane grows as a function of stimulus level. An example of this is shown in Figure 1.2, where the velocity of basilar membrane movement is plotted as a function of stimulus level. This type of function is generally referred to as a basilar membrane input-output function. These results are from the basal region of a chinchilla cochlea, a region that responds best to high frequencies. The CF of this particular recording site was 10,000 Hz. The results in Figure 1.2 are in response to a 10,000-Hz tone (i.e., a tone at CF). The magnitude of the response generally increases with increasing stimulus level, but the growth is quite compressive. This is clear by comparing the input-output function to the linear function (dashed line). Throughout its most compressive region (at moderate to high stimulus levels), the input-output function has a slope of about 0.2 dB/dB, which corresponds to a compression ratio of about 5 : 1. Interestingly, the input-output function is compressive only for stimulus frequencies near the CF of the recording site. The input-output function is linear for stimulus frequencies well below or well above the CF (e.g., Ruggero et al. 1997; Cooper, Chapter 2). As a result of this compression, any given point along the basilar membrane is able to respond to an extremely large range of stimulus levels. Because basilar membrane motion serves as the proximal stimulus for the inner hair cells and, subsequently, the auditory nerve fibers, the compression that is observed at the basilar membrane greatly extends the dynamic range of the peripheral auditory system. 2 The CF for a given place is typically defined as the frequency that requires the least sound pressure level to generate a criterion response; thus it is defined using relatively low stimulus levels. At high sound pressure levels, a given place may respond best to a frequency lower than its nominal CF (see Cooper, Chapter 2). This corresponds to a basal shift in the peak of the traveling wave with increasing stimulus level (e.g., Ruggero et al. 1997; Rhode and Recio 2000). Throughout this chapter, CF refers to the frequency that is most effective at eliciting a response at a given place in the cochlea at relatively low levels.

1. Overview of Auditory Compression

5

Figure 1.2. An example of a basilar membrane input-output function (circles connected by a solid line). The stimulus frequency (10,000 Hz) was equal to the characteristic frequency (CF) of the recording site. Dashed line: linear growth of response. Note that the input-output function is highly compressive. (Data from Ruggero et al. 1997.)

2.1 The Role of Outer Hair Cells Although compression can be measured at the basilar membrane, it is not due to the mechanics of the membrane per se. In other words, the basilar membrane by itself is not compressive. Where does this compression come from? Some insight into that question can be gained from the results in Figure 1.3 that show basilar membrane input-output functions measured at the 11,000-Hz place both before and after an intravenous injection of the drug quinine. Quinine is ototoxic and is thought to directly affect the outer hair cells in the cochlea (Karlsson and Flock 1990; Jarboe and Hallworth 1999; Zheng et al. 2001). In terms of the response to a tone at CF, quinine reduces the magnitude of the basilar membrane response at the lower stimulus levels but not at the higher levels. In other words, a drug that adversely affects the outer hair cells can result in a smaller mechanical response at the basilar membrane. Because the effect of quinine is relatively large at low levels and essentially nonexistent at high levels, the consequence of this temporary outer hair cell damage is a more linear input-output function. Quinine has no effect on the response to a relatively low-frequency tone (data not shown) where the growth of response at the basilar membrane is

6

S.P. Bacon

Figure 1.3. The effects of quinine on the basilar membrane input-output function for a tone at which the frequency (11,000 Hz) was equal to the CF of the recording site. The input-output function was measured before and after administration of quinine. (Data are from Recio and Ruggero 1995.)

already linear (see Fig. 2.9 in Ruggero et al. 1996). This suggests that the outer hair cells only affect the response to a tone at (or near) the CF where the growth is normally compressive. A variety of manipulations that have resulted in various degrees of temporary or permanent hearing loss have yielded results similar to those described above. The main finding is summarized schematically in Figure 1.4. For a tone at CF, the growth of response is compressive under normal conditions but is linear (or at least more linear) when the outer hair cells are damaged or functioning abnormally. Although the precise way in which the outer hair cells affect the motion of the basilar membrane is unclear, it is likely the result of outer hair cell electromotility. Indeed, a fascinating finding from research on cochlear hair cells is that the outer hair cells have motor capability resulting in their being motile and, in isolation, being capable of changing shape at rates in the audio frequency range (Brownell et al. 1985; Kachar et al. 1986; Zenner 1986; Ashmore 1987; for a review, see Holley 1996). Recently, Zheng et al. (2000) have identified the motor protein (prestin) responsible for this electromotility. These shape changes are thought to alter the micromechanical properties of the cochlea so as to increase the vibration of the basilar membrane in a frequency-selective way. In other words, they provide local mechanical amplification in the form of feedback. For this reason, the outer hair cells are often referred to as the “cochlear amplifier” (Davis 1983; for a review, see Robles and Ruggero

1. Overview of Auditory Compression

7

Figure 1.4. Summary of the effects of outer hair cell (OHC) damage on the basilar membrane input-output function for a tone at the CF of the recording site. Under normal conditions, the function is compressive. After outer hair cell damage, the function becomes linear (or at least more linear) and there is a concomitant loss of gain, particularly at lower SPLs.

2001). Damage to the outer hair cells results in a loss of that amplification or gain, as shown in Figure 1.4. The amount of gain that normally exists has been estimated to be as large as 50–80 dB for lower sound pressure levels (Ruggero et al. 1997). The gain decreases with increasing level and is negligible at high levels (as evidenced by the horizontal difference between the solid and dashed lines in Fig. 1.4). This amplification is responsible for our incredible sensitivity to sound. That is, this cochlear amplification is what enables us to detect low-level sounds, thereby lending strong support to the claim that the outer hair cells are largely responsible for our enormous dynamic range of hearing.

3. Some Perceptual Consequences of Basilar Membrane Compression An important goal in hearing science is to explain auditory perceptual phenomena in terms of underlying physiological mechanisms. Although the following point may be obvious, it is nevertheless worth noting that psychophysical measurements do not reflect solely the response of any one part of the auditory system but instead reflect the response of the entire system. This clearly makes it difficult to conclude with absolute certainty that a particular psychophysical phenomenon reflects largely or primarily the response of a specific site in the system. Nevertheless, it is possible to

8

S.P. Bacon

observe psychophysical correlates of processing that exist at a specific site in the auditory system. Furthermore, the greater the similarity between the psychophysical and physiological observations or measurements, the stronger the putative connection between the two. Sufficient evidence exists linking our exquisite sensitivity to functioning outer hair cells that few would argue against the claim that the extremely large dynamic range of hearing is one perceptual consequence of basilar membrane compression. There are many other perceptual phenomena that likewise appear to be closely linked to basilar membrane compression. Oxenham and Bacon (Chapter 3) discuss these in detail, whereas the remainder of this section provides an overview of just a few such phenomena. Masking experiments have been used for well over a century to gain insight into the way in which the auditory system processes sound (Mayer 1876). The focus often has been on the spectral characteristics of masking so that these masking experiments have concentrated primarily on revealing the limits of our frequency-resolving (or filtering) capabilities. It is likely that the filtering that exists along the basilar membrane is reflected in the psychophysical estimates of auditory filtering (Fletcher 1940; Moore 1993; Shera et al. 2002). Indeed, it has often been argued that the auditory filter (or critical band) corresponds to a certain distance along the basilar membrane (Fletcher 1940; Zwicker et al. 1957; Greenwood 1961, 1990; Moore 1986; Glasberg and Moore 1990). This distance is usually estimated at about 0.9 mm. In other words, the auditory filter bandwidth is thought to correspond to a distance along the basilar membrane of about 0.9 mm. Given the apparently close tie between psychophysical masking and basilar membrane mechanics, it is perhaps not surprising that the growth of masking might be intimately linked to the growth of response at the basilar membrane. It seems likely, for example, that the well-known “upward spread of masking” (e.g., Wegel and Lane 1924; Egan and Hake 1950; see Fig. 3.2 in Oxenham and Bacon, Chapter 3) can be explained by basilar membrane compression. This is where the masking produced by a narrowband masker spreads considerably more to higher frequencies than it does to lower frequencies, growing expansively at the higher frequencies as the masker level increases to high sound pressure levels. Moreover, it is now believed that the rate at which this upward spread of masking grows with masker level can provide an estimate of the degree of compression measured at the basilar membrane (Oxenham and Plack 1997; Oxenham and Bacon, Chapter 3; Bacon and Oxenham, Chapter 4), at least when the masker and signal do not overlap in time, such as in forward masking. To the extent to which this is true, these behavioral experiments can be used to obtain noninvasive estimates of the growth of basilar membrane response and, by implication, insight into the functioning of the outer hair cells in the cochlea. Whether these types of measures will ever make their way into the audiology clinic as a diagnostic test of outer hair cell function remains to be seen, but their utility in the laboratory is clear.

1. Overview of Auditory Compression

9

The apparent influence of basilar membrane compression can also be seen in a variety of measures of temporal processing such as forward masking and the masking by temporally fluctuating maskers. This is particularly interesting given that temporal measures are often thought to reflect processing primarily at a site in the auditory system that is central to the auditory nerve. Of course, peripheral processing is obligatory and could obviously have an effect on any measure, even those that reflect significant amounts of central processing. As discussed by Oxenham and Bacon (Chapter 3), the role of basilar membrane compression in these types of measures has been evaluated within the context of a temporal window model (see Fig. 3.5 in Oxenham and Bacon, Chapter 3). This model incorporates a static nonlinearity that is thought to represent basilar membrane compression. The degree of compression that is needed to account for various psychophysical data is generally consistent with the amount of compression observed directly via mechanical measurements at the basilar membrane. Furthermore, the results from individuals with cochlear hearing loss (Bacon and Oxenham, Chapter 4) often can be predicted rather well with the complete elimination of compression in the model, providing converging evidence in support of an important role for basilar membrane compression in these measures.

4. Some Perceptual Consequences of Reduced or Absent Compression Most individuals with a permanent sensorineural hearing loss suffer from damage to the outer hair cells in the cochlea. As discussed in Section 2.1, this damage results in a reduction in or loss of basilar membrane compression. Thus these individuals provide an opportunity to evaluate the potential role of basilar membrane compression in hearing. In addition, it is possible to adversely affect outer hair cells temporarily via well-controlled exposures to intense sound or the ingestion of moderately high doses of aspirin. Individuals with a temporary loss induced by either of these agents provide an especially interesting and informative subject population for studying the role of basilar membrane compression.This is because the temporary loss is restricted to the outer hair cells, whereas with a permanent loss, the inner hair cells may be damaged as well. Bacon and Oxenham (Chapter 4) discuss research on individuals with either a temporary or permanent cochlear hearing loss and evaluate the extent to which the various perceptual consequences of cochlear damage can be understood in terms of a loss of basilar membrane compression. An example of some of that research is highlighted below. As noted in Section 3, the growth of masking may provide an estimate of the growth of response at the basilar membrane. Supporting evidence

10

S.P. Bacon

for that comes from masking experiments in individuals with a temporary, aspirin-induced hearing loss (Hicks and Bacon 1999, 2000). Some results are shown in Figure 1.5. In these types of experiments, the masker is usually much lower in frequency than the signal so that the response to the masker at the signal frequency place (where the subject is likely to detect the signal) will be linear, whereas the response to the signal at that same place will be compressive (for more details, see Oxenham and Bacon, Chapter 3; Bacon and Oxenham, Chapter 4). In this particular experiment, the signal was fixed in level and the masker was varied in level to mask the signal. The experimental paradigm was forward masking, and thus the brief signal occurred immediately after the offset of the masker. The slope of the masking function is more gradual (i.e., more compressive) before aspirin ingestion than it is after aspirin ingestion. This indicates that the aspirin has reduced the amount of compression in the auditory system. Note that aspirin has its greatest effect at lower signal levels and has little or no effect at higher levels. Importantly, this mimics the effect of quinine on the growth of response at the basilar membrane (see Fig. 1.3), thus strengthening the argu-

Figure 1.5. An example of the effect of aspirin on growth-of-masking functions in forward masking. The frequency of the masker (2,222 Hz) was well below the frequency of the signal (4,000 Hz). The masking function was measured before and after administration of aspirin, which resulted in a temporary hearing loss of about 13 dB. Dashed line: linear growth. (Data are from Hicks and Bacon 2000.)

1. Overview of Auditory Compression

11

ment that the behavioral experiments estimate the growth of response measured directly at the basilar membrane. Similar masking experiments have been used to examine how permanent sensorineural hearing loss affects compression (Oxenham and Plack 1997; Nelson et al. 2001). These results suggest that relatively severe, permanent loss can completely eliminate the compressive growth of response. As discussed in Section 2.1, the reduction in or loss of basilar membrane compression as a consequence of outer hair cell damage is accompanied by a decrease in or loss of gain that is normally observed at low stimulus levels. The most obvious perceptual consequence of this damage is an elevation of the absolute threshold for sound, which represents the hearing loss. Because the level at which sounds become uncomfortably loud usually does not change much with hearing loss, the elevated absolute thresholds result in a reduced dynamic range of hearing. For individuals with a cochlear hearing loss, the loudness of a sound goes from relatively soft to uncomfortably loud over a smaller range of sound pressure levels than it does in individuals with normal hearing. This is referred to as “loudness recruitment.” The reduced dynamic range in individuals with a cochlear hearing loss poses a problem in terms of rehabilitation: a large range of stimulus levels must be “squeezed” or compressed into the relatively small dynamic range.

5. Hearing Aids The most common form of rehabilitation for individuals with a cochlear hearing loss is amplification. The intent is to amplify sounds so that the individual with the hearing loss can hear them. The reduced dynamic range, however, provides a considerable challenge. This is illustrated in Figure 1.6. In Figure 1.6A, the growth of loudness is shown schematically for a normal (solid line) and an impaired (dashed line) ear. The specific goal of a hearing aid might be to amplify low-level sounds a great deal but high-level sounds only a little, if at all, so as to shift the response of the impaired ear to be more in line with the response of the normal ear (as indicated by the arrows). As noted in Section 2.1, outer hair cells normally provide this type of level-dependent amplification. That is, the amount by which they amplify the vibration of the basilar membrane decreases with increasing input level. This goal of level-dependent amplification will not be accomplished by simple linear amplification where all sounds are amplified by the same amount. Instead, compression amplification has become an increasingly more popular type of amplification for individuals with cochlear hearing loss in order to deal successfully with their reduced dynamic range. An illustration of one type of compression amplification is shown in Figure 1.6B, which shows an example of the input-output function of a compression hearing aid. In this case, the function is linear up to an input level of 40 dB

12

S.P. Bacon

Figure 1.6. A: growth of loudness (in sones) for a normal (solid line) and an impaired (dashed line) ear. The length of the arrows indicates the amount by which the sound would need to be amplified to shift the impaired ear to be equal to the normal ear. B: input-output function for a hearing aid that is linear at input levels of 40 dB SPL and below and compressive at levels above that. C: gain provided by that hearing aid.

and compressive above that. The gain (the difference between the output and the input) for this function is shown in Figure 1.6C. Note that this type of amplification has accomplished the specific goal of providing considerable gain at low levels and increasingly smaller amounts of gain at higher levels. In a broad sense, compression hearing aids are attempting to “restore” the compression that is normally observed and, consequently, extend the dynamic range of people with hearing loss. As discussed by

1. Overview of Auditory Compression

13

Levitt (Chapter 5), hearing aids can provide various degrees of compression and can provide different amounts of compression in different frequency regions in order to accommodate different severities and configurations of hearing loss.

6. Cochlear Implants Some individuals have a hearing loss that is so severe that they will not benefit from acoustic amplification. Those individuals are often fitted with a cochlear implant, which bypasses the damaged cochlea and stimulates the auditory nerve directly with electric current. The dynamic range of individuals with a cochlear implant is typically on the order of only about 5–20 dB in electric current, which is much smaller than the typical acoustic dynamic range of a hearing-impaired individual fitted with a hearing aid. Thus the amount of compression required to map the acoustic amplitude to electric current may be even greater than the compression required in typical hearing aids. Although cochlear implants are relatively new, a considerable amount of psychophysical research has been conducted in individuals fitted with cochlear implants over the past 20 years. Some of that research has focused on determining the appropriate compression required in the signal-processing stage to go from acoustic amplitude to electric current. Zeng (Chapter 6) summarizes this research and discusses, among other things, the important issues faced by scientists and clinicians when trying to map the large range of acoustic levels in the auditory environment to an auditory system that does not benefit from the compression that normally exists at the basilar membrane in the cochlea.

7. Summary One function of the cochlea is to transduce mechanical vibrations along the basilar membrane into neural impulses that can be processed by the auditory nervous system. The auditory nerve fibers that convey neural information from the cochlea synapse directly with the inner hair cells. Fibers that synapse with the outer hair cells are apparently unresponsive to sound and thus do not contribute directly to the neural code. Nevertheless, research over the last 20 or so years has clearly elucidated the importance of outer hair cells for hearing. Amazingly, these sensory cells have a motor function.As a result of their motile properties, they affect basilar membrane mechanics and hence, ultimately, the neural information conveyed by the auditory nerve fibers. The outer hair cells enhance or amplify the vibration of the basilar membrane in a frequency-selective way at low stimulus levels, but the amount of that amplification decreases with increasing stimulus level, leading to a compressive growth of response at the basilar membrane.

14

S.P. Bacon

The amplification accounts for the exquisite sensitivity of the auditory system and, consequently, its impressively large dynamic range of hearing. And because the amplification is frequency selective, it is responsible for the excellent frequency-resolving (or filtering) capabilities of the system. Psychophysical research in subjects with normal hearing and subjects with either temporary or permanent cochlear hearing loss, together with various modeling efforts, suggests that the perceptual consequences of basilar membrane compression may be far-reaching. Temporary or permanent damage to the outer hair cells in the cochlea results in a decrease in or loss of the amplification provided by those cells and hence a decrease in or loss of basilar membrane compression (among other things). The perceptual consequences of this damage are significant and lie at the heart of the auditory processing difficulties experienced by hearing-impaired individuals. Hearing aids and cochlear implants are important rehabilitative devices for individuals with a cochlear hearing loss. These devices must cope with the loss of basilar membrane compression and the subsequent reduction in the dynamic range of hearing. By incorporating compression in their signal-processing schemes, they can map the large level variations in the acoustic environment into the significantly reduced dynamic range of hearing experienced by these patients. In a sense, they restore some of the important compression that normally exists at the basilar membrane of the cochlea. The importance of peripheral auditory compression is becoming increasingly clear as auditory scientists from various disciplines focus on the role of the outer hair cells in hearing. Considerable advances in understanding the nature and perceptual consequences of basilar membrane compression have occurred primarily over the past two decades. It is fair to say that future research will provide important new insights into this compression and the extent to which it influences auditory behavior. Some of this future research undoubtedly will focus on whether the degree and nature of compression vary along the length of the cochlea from base to apex. Most of the direct measurements of basilar membrane motion to date have been restricted to the basal region. There is, however, some evidence that there is less compression at the apical region and that the compression there is less frequency selective (see Cooper, Chapter 2). Similarly, there is psychophysical evidence that the degree of compression is greatly reduced at low frequencies, although the possibility that compression may be less frequency selective at low frequencies complicates most psychophysical estimates of compression at those frequencies. This, in turn, complicates the comparisons of compression across frequency (see Oxenham and Bacon, Chapter 3). Research on compression also might provide important insight into why hearing-impaired individuals with similar amounts of hearing loss often perform quite differently on a range of auditory tasks. One possibility is that the differences in performance might be explained by differences in

1. Overview of Auditory Compression

15

the degree of residual compression and, by implication, differences in the amount of residual outer hair cell function. Advances in the rehabilitation of cochlear hearing loss will also benefit from additional research on compression. In this regard, the important issues for both hearing aids and cochlear implants revolve around determining the optimum compression settings that will maximize the understanding of speech while at the same time providing the listener with a high level of overall satisfaction with the device. These and many other issues will be at the forefront of future research as scientists continue to strive not only to understand the complex and pervasive role that basilar membrane compression plays in hearing but also to determine the best ways to compensate for the reduction in or loss of compression that accompanies cochlear hearing loss.

Acknowledgments. The writing of this chapter was partially supported by National Institute on Deafness and Other Communication Disorders Grant DC-01376. I thank Li Liu for her assistance with the figures.

References Ashmore JF (1987) A fast motile response in guinea-pig outer hair cells: the cellular basis of the cochlear amplifier. J Physiol (Lond) 388:323–347. Brownell WE, Bader CR, Bertrand D, de Ribaupierre Y (1985) Evoked mechanical responses of isolated cochlear outer hair cells. Science 227:194– 196. Dallos P (1988) Cochlear neurobiology: some key experiments and concepts of the past two decades. In: Edelman GM, Gall WE, Cowan WM (eds) Auditory Function: Neurobiological Bases of Hearing. New York: John Wiley and Sons, pp. 153–189. Dallos P (1992) The active cochlea. J Neurosci 12:4575–4585. Davis H (1983) An active process in cochlear mechanics. Hear Res 9:79–90. Delgutte B (1996) Physiological models for basic auditory percepts. In: Hawkins HL, McMullen TA, Popper AN, Fay RR (eds) Auditory Computation. New York: Springer-Verlag, pp. 157–220. Egan JP, Hake HW (1950) On the masking pattern of a simple auditory stimulus. J Acoust Soc Am 22:622–630. Evans EF (1981) The dynamic range problem: place and time coding at the level of cochlear nerve and nucleus. In: Syka J, Aitkin L (eds) Neuronal Mechanisms in Hearing. New York: Plenum Press, pp. 69–85. Fletcher H (1940) Auditory patterns. Rev Mod Phys 12:47–65. Glasberg BR, Moore BCJ (1990) Derivation of auditory filter shapes from notchednoise data. Hear Res 47:103–138. Greenwood DD (1961) Critical bandwidth and the frequency coordinates of the basilar membrane. J Acoust Soc Am 33:1344–1356. Greenwood DD (1990) A cochlear frequency-position function for several species— 29 years later. J Acoust Soc Am 87:2592–2605.

16

S.P. Bacon

Hicks ML, Bacon SP (1999) Effects of aspirin on psychophysical measures of frequency selectivity, two-tone suppression, and growth of masking. J Acoust Soc Am 106:1436–1451. Hicks ML, Bacon SP (2000) The effects of aspirin on a psychophysical estimate of basilar membrane compression. J Acoust Soc Am 107:2914. Holley MC (1996) Outer hair cell motility. In: Dallos P, Popper AN, Fay RR (eds) The Cochlea. New York: Springer-Verlag, pp. 386–434. Jarboe JK, Hallworth R (1999) The effect of quinine on outer hair cell shape, compliance and force. Hear Res 132:43–50. Kachar B, Brownell WE, Altschuler R, Fex J (1986) Electrokinetic shape changes of cochlear outer hair cells. Nature 322:365–368. Karlsson KK, Flock Å (1990) Quinine causes isolated outer hair cells to change length. Neurosci Lett 116:101–105. Kros CJ (1996) Physiology of mammalian cochlear hair cells. In: Dallos P, Popper AN, Fay RR (eds) The Cochlea. New York: Springer-Verlag, pp. 318–385. Levitt H (1982) Speech discrimination ability in the hearing impaired: spectrum considerations. In: Studebaker GA, Bess FH (eds) The Vanderbilt Hearing Aid Report (Monographs in Contemporary Audiology). Upper Darby, Pa, pp. 32–43. Liberman MC (1982) Single-neuron labeling in the cat auditory nerve. Science 216:1239–1241. Liberman MC, Oliver ME (1984) Morphometry of intracellularly labeled neurons of the auditory nerve: correlations with functional properties. J Comp Neurol 223: 163–176. Mayer AM (1876) Researches in acoustics. Philos Mag 2:500–507. Moore BCJ (1986) Parallels between frequency selectivity measured psychophysically and in cochlear mechanics. Scand Audiol Suppl 25:139–152. Moore BCJ (1993) Frequency analysis and pitch perception. In: Yost WA, Popper AN, Fay RR (eds) Human Psychophysics. New York: Springer-Verlag, pp. 56–115. Moore BCJ (1998) Cochlear Hearing Loss. London: Whurr Publishers. Nelson DA, Schroder AC, Wojtczak M (2001) A new procedure for measuring peripheral compression in normal-hearing and hearing-impaired listeners. J Acoust Soc Am 110:2045–2064. Oxenham AJ, Plack CJ (1997) A behavioral measure of basilar-membrane nonlinearity in listeners with normal and impaired hearing. J Acoust Soc Am 101:3666–3675. Patuzzi R (1996) Cochlear micromechanics and macromechanics. In: Dallos P, Popper AN, Fay RR (eds) The Cochlea. New York: Springer-Verlag, pp. 186–257. Recio A, Ruggero MA (1995) Effects of quinine on basilar-membrane responses to sound. Assoc Res Otolaryngol Midwinter Mtg Abstr 18:200. Rhode WS, Recio A (2000) Study of mechanical motions in the basal region of the chinchilla cochlea. J Acoust Soc Am 107:3317–3332. Robertson D (1984) Horseradish peroxidase injection of physiologically characterised afferent and efferent neurons in the guinea pig spiral ganglion. Hear Res 15:113–121. Robles L, Ruggero MA (2001) Mechanics of the mammalian cochlea. Physiol Rev 81:1305–1352. Ruggero MA (1992) Physiology and coding of sound in the auditory nerve. In: Popper AN, Fay RR (eds) The Mammalian Auditory Pathway: Neurophysiology. New York: Springer-Verlag, pp. 34–93.

1. Overview of Auditory Compression

17

Ruggero MA, Rich NC, Robles L, Recio A (1996) The effects of acoustic trauma, other cochlear injury, and death on basilar-membrane responses to sound. In: Axelsson A, Borchgrevink H, Hamernik RP, Hellström P, Henderson D, Salvi RJ (eds) Scientific Basis of Noise-Induced Hearing Loss. New York: Thieme, pp. 23–35. Ruggero MA, Rich NC, Recio A, Narayan SS, Robles L (1997) Basilar-membrane responses to tones at the base of the chinchilla cochlea. J Acoust Soc Am 101:2151–2163. Shera CA, Guinan JJ Jr, Oxenham AJ (2002) Revised estimates of human cochlear tuning from otoacoustic and behavioral measurements. Proc Natl Acad Sci USA 99:3318–3323. Slepecky NB (1996) Structure of the mammalian cochlea. In: Dallos P, Popper AN, Fay RR (eds) The Cochlea. New York: Springer-Verlag, pp. 44–129. Viemeister NF (1988a) Intensity coding and the dynamic range problem. Hear Res 34:267–274. Viemeister NF (1988b) Psychophysical aspects of auditory intensity coding. In: Edelman GM, Gall WE, Cowan WM (eds) Auditory Function: Neurobiological Bases of Hearing. New York: John Wiley and Sons, pp. 213–242. Viemeister NF, Bacon SP (1988) Intensity discrimination, increment detection, and magnitude estimation for 1-kHz tones. J Acoust Soc Am 84:172–178. von Békésy G (1960) Experiments In Hearing. New York: McGraw-Hill. Wegel RL, Lane CE (1924) The auditory masking of one sound by another and its probable relation to the dynamics of the inner ear. Phys Rev 23:266–285. Zenner HP (1986) Motile responses in outer hair cells. Hear Res 22:83–90. Zheng J, Shen W, He DZ, Long KB, Madison LD, Dallos P (2000) Prestin is the motor protein of cochlear outer hair cells. Nature 405:149–155. Zheng J, Ren T, Parthasarathi A, Nuttall AL (2001) Quinine-induced alterations of electrically evoked otoacoustic emissions and cochlear potentials in guinea pigs. Hear Res 154:124–134. Zwicker E, Flottorp G, Stevens SS (1957) Critical band width in loudness summation. J Acoust Soc Am 29:548–557.

2 Compression in the Peripheral Auditory System Nigel P. Cooper

1. Introduction 1.1 Overview The peripheral auditory system converts sound into an informative ensemble of neural signals. The range of sounds that it deals with is immense. Most mammals can accurately distinguish spectral features spanning at least two decades of frequency and six decades of intensity. To a large extent, this performance is made possible through the biologically and physically active processing of sound that occurs in the cochlea (the acoustic part of the inner ear). The purpose of this chapter is to describe this processing in some detail, with particular emphasis on the coding of the intensity of a sound. The coding of the spectrum of a sound, through cochlear frequency analysis, is the subject of detailed reviews elsewhere (e.g., Patuzzi and Robertson 1988; Patuzzi 1996). As far as intensity coding is concerned, the primary function of the peripheral auditory system is to combine high sensitivity with a large dynamic range. It achieves this by combining amplification with compression. For the purposes of this review, it is important to make the meanings of these two terms very clear. Amplification simply means “making something larger,” whereas compression means “fitting more into an available space than would otherwise be possible.” The two terms are not reciprocal; hence something that is more compressed is not necessarily less amplified, and something that is more amplified is not, of necessity, less compressed. The idea that amplification and compression are reciprocal is a common misunderstanding, and one that can lead to much confusion when interpreting data from the peripheral auditory system. As we shall see, a healthy peripheral auditory system fits a lot more information into a limited amount of output space by the selective amplification of its responses to low-level stimuli.

18

2. Peripheral Compression

19

1.2 A Brief Review of the Anatomy and Physiology of the Peripheral Auditory System The peripheral auditory system comprises the external ear, the middle ear, the cochlea, and the auditory nerve (the acoustic part of the eighth cranial nerve). The external and middle ears are primarily concerned with the efficient transmission of sound from an animal’s environment to its inner ear. This transmission is generally regarded to be a linear process, such that the spectral and temporal features of the sound are preserved with high fidelity at the input to the cochlea (Guinan and Peake 1967; see Rosowski 1994 for a review). However, the middle ear does afford some means of amplitude control on the transmitted sounds. Two of the small, articulated bones within the middle ear are attached to a pair of muscles known as the tensor tympani and the stapedius muscles. The bones in question are the malleus and the stapes, which connect to the eardrum and the oval window of the cochlea, respectively. The muscles can alter the position of both of these bones as well as the tension under which they operate. The muscles can therefore alter the acoustic impedance1 of the middle ear, and this, in turn, affects the efficiency of a sounds transmission to the cochlea. The middle ear muscles are normally activated only in response to very loud sounds, and their primary function is therefore commonly considered to be protective (Borg et al. 1984; Pang and Peake 1986). However, the muscles can also be activated voluntarily, and they are thought to contract routinely during vocalization (Borg and Zakrisson 1975). The muscles reduce the transmission of low-frequency sounds more than high-frequency sounds, so they can often be used to improve the detection of signals in background noise (Pang and Guinan 1997). Contraction of the middle ear muscles is also known to improve the intelligibility of speech at high sound pressure levels (SPLs) (Borg and Counter 1989). Because the middle ear muscles effectively operate by turning down the volume of high-level sounds, they can be considered as one means of achieving compression in the peripheral auditory system. When compared to the amount of compression to be achieved within the cochlea, however, their significance is lessened considerably. The anatomical structure of the mammalian cochlea is fairly complex, and has been described in great detail elsewhere (e.g., see Slepecky 1996). For the purposes of this chapter, the gross structure of the cochlea can be simplified into the form illustrated in Figure 2.1A. All of the space within the cochlea is filled with fluids. These fluids are split into three chambers or scalae by the structures of the cochlear partition (only two of the chambers are evident in Fig. 2.1A, but all three are shown in Fig. 2.1B,C). The fluids in two of the chambers, the scala vestibuli and the scala tympani, can 1

The acoustic impedance of a material determines how much sound pressure is needed to cause a particular amount of movement within that material.

20

N.P. Cooper apex

basilar membrane

stapes

oval window

uli scala vestib helicotrema base scala

i tympan

cochlear partition

A

round window Reissner's membrane

laser apex

chamber

scala vestibuli

4

tectorial membrane IHC

3 1mm

OHCs

organ of Corti scala media

2 base

bone

1

B

auditory nerve

cover glass laser

basilar membrane

C

osseous spiral lamina ANFs

spiral ligament scala tympani

Figure 2.1. Functional anatomy of the mammalian cochlea. A: layout of a hypothetically straightened or uncoiled cochlea. The cochlear partition is represented as a flat plate that separates the fluid-filled spaces of the cochlea into 2 main chambers. B: schematic cross section through a real mammalian cochlea, showing 2 experimental approaches to the cochlear partition.The diagram is based on a cross section of a guinea pig cochlea in which the cochlear partition spirals four times around the trunk of the auditory nerve. C: schematic cross section through the apical turn of a real cochlea showing some of the cellular detail of the partition. IHC, inner hair cells; OHCs, outer hair cells; ANFs, auditory nerve fibers. (B and C adapted with permission from Cooper 1999b. Copyright © 1999 Elsevier Science.)

communicate with one another through the helicotrema, a small hole close to one end of the cochlear partition. The fluids can also communicate with the outside world through the two “windows” in the cochlea. The oval window is closely apposed to the stapes of the middle ear and allows sounds to be transferred into the cochlea very efficiently. The round window, on the other hand, vents the pressure in the cochlear fluid back to that in the airfilled cavities of the middle ear. This arrangement means that the middle ear does not have to work too hard to pressurize the fluid-filled inner ear in order to transmit a sound into it; it merely has to move the fluid back and forth in synchrony with the sound (see Patuzzi 1996). When the stapes pushes the oval window into (out of) the cochlea, the fluid moves almost instantaneously along the cochlear ducts and the round window bulges outward (inward) to preserve the volume of the fluid. The movements of

2. Peripheral Compression

21

the fluid are driven by the pressure gradient that exists between the oval and round windows. The details of this gradient are determined by the flexible parts of the cochlear partition, which act to set up a secondary series of pressure waves that travel along the length of the cochlea (see below). The difference between the pressures on the two sides of the cochlear partition (between the scala vestibuli and scala media on the one side and the scala tympani on the other) forces the flexible parts of the partition (including the basilar membrane, the organ of Corti, the tectorial membrane, and Reissner’s membrane; see Fig. 2.1C) to move up and down in synchrony with the sounds (see black and white half-arrows in Fig. 2.1A,C). Not all of the cross sections of the partition move at once, however. The time that the “traveling” pressure waves take to reach a particular cross section depends on the position of the cross section along the length of the cochlea as well as both the frequency and intensity of the sound. In general, both the pressure waves and the vibration waves that they set up tend to travel from the base to the apex of the cochlea, with the low-frequency components of the waves traveling further and slightly faster than the highfrequency components, and the components of the low-intensity sounds traveling slightly further and slightly more slowly than those of the higher intensity sounds (for reviews see Davis 1983; Patuzzi 1996). The mapping of the spectral characteristics of a sound onto the spatial coordinates of the cochlear partition is known as tonotopy, and is a feature of the fundamental importance to the performance of the entire auditory system. It is brought about by variations in the physical characteristics of the cochlear partition, such as its stiffness, its mass, and its architecture. All of these characteristics change systematically from one end of the cochlea to the other (Fernàndez 1952; von Békésy 1960; Lim 1980; see Patuzzi 1996 for a review). The structures that actually detect the presence of a sound within the cochlea are a group of highly specialized receptor cells known as hair cells. These cells can be subdivided into two groups depending on their location across the width of the cochlear partition (see Fig. 2.1C): The inner hair cells line up in a single row at the edge of the organ of Corti, directly above the border between the flexible basilar membrane and the inflexible osseous spiral lamina (see Fig. 2.1C), whereas the outer hair cells are distributed across three (or occasionally four) rows in the center of the organ of Corti, directly above the most flexible region of the basilar membrane. The inner hair cells are heavily innervated by the afferent fibers that form the auditory nerve, and their main function is to signal the presence of a sound to the central nervous system (Spoendlin 1967). In stark contrast, the outer hair cells receive relatively little afferent innervation, but are the targets of numerous efferent nerve fibers that run from the brainstem to the organ of Corti (see Guinan 1996 for a review). The main function of the outer hair cells appears to be to select, amplify, and compress the acoustic signals that actually reach the inner hair cells. They do this by affecting the mechanics of the cochlear partition.

22

N.P. Cooper

Despite the clear division of labor that exists between the two types of hair cells in the mammalian cochlea, the mechanism by which each cell transduces a sound is essentially the same. Each hair cell possesses many tens, or possibly hundreds, of mechanically sensitive transducer channels at the tips of a highly specialized bundle of “hairs” or stereocilia. The stereocilia project from the uppermost surfaces of the hair cells and respond to stimuli that deflect their tips sideways in the plane in Figure 2.1C (cf. Hudspeth and Corey 1977). The conversion of a sound into such a stimulus is a result of the peculiar architecture of the organ of Corti: upward (downward) motion of the basilar membrane causes the innermost parts of the organ of Corti to rotate about their attachment to the osseous spiral lamina and leads to inward (outward) motion of the bases of the stereocilia with respect to the overlying tectorial membrane (cf. ter Kuile 1900). The amount of movement that is necessary to change the electrical properties of the transducer channels in the stereocilia of each hair cell is miniscule; displacements of less than ±100 nm are sufficient to change almost all of the channels in a given hair cell from an open, highly conductive state into a closed, nonconductive state (see Kros 1996). The state changes of the channels alter the electrical conductivity of the hair cells and give rise to sound-evoked receptor potentials between the inside and outside of the hair cells (Russell and Sellick 1978; Dallos 1985). In the case of the inner hair cells, the receptor potentials can trigger or modulate the release of chemical neurotransmitters onto the innervating fibers of the auditory nerve (Palmer and Russell 1986; Siegel 1992). In the case of the outer hair cells, the receptor potentials affect the mechanical properties of the basolateral wall of the cells, leading to changes in the length of the cells (Brownell et al. 1985; Ashmore 1987; Zheng et al. 2000) and the mechanical properties of the entire cochlear partition (see de Boer 1996; Holley 1996; Patuzzi 1996 for reviews). The mechanical sensitivity of the hair cells in the mammalian cochlea appears to have evolved at a considerable cost: as well as being remarkably sensitive, the transduction mechanisms are incredibly fragile and highly nonlinear. The fragility of the mechanism is thought to be the source of many forms of hearing loss (see Patuzzi et al. 1989) and can lead to all sorts of problems in experimental studies of cochlear function. The nonlinearity is a consequence of the molecular nature of the transducer (see Kros 1996 for a review) and is thought to be the source of almost all of the compression that occurs in the cochlea (see Section 2.2).

2. Compression in the Mechanics of the Cochlea 2.1 Observations Direct observations of the cochlea’s mechanical responses to sound have provided the most detailed, wide-ranging, and readily interpretable infor-

2. Peripheral Compression

23

mation that is available to date on the intensity-coding performance of the peripheral auditory system. These observations have been made by numerous investigators using numerous techniques over numerous years, and it is not the intention of this chapter to review them all. The rationale behind the exclusion of many observations is quite straightforward: direct observations made in early mechanical studies revealed very little about the way in which we now believe a normal cochlea processes the intensity of a sound because they were made using techniques that caused too much trauma to the extremely vulnerable cochlea. The first observations to have overcome the problems of trauma were only provided in the 1970s, and these were neither confirmed nor commonly accepted until the early 1980s (for historical reviews, see Patuzzi 1996; Robles and Ruggero 2001). Most of the observations to be described in this section have hence been made in the last two decades. There are two major limitations to all of the direct observations of cochlear mechanics that have been made to date. The first is that the observations have been limited to only a few locations within the cochlea. This is a straightforward consequence of the peculiar anatomy of the cochlea. Instead of having their fluid-filled ducts arranged in a straight line, as shown in Figure 2.1A, real mammalian cochleae consist of tightly coiled helices that are at least partially embedded in one of the hardest bones of the body (cf. Fig. 2.1B). Making observations from even the most accessible parts of the cochlea therefore involves fairly major surgery, which often leads to trauma and pathophysiological conditions. The way that most investigators get around this problem is to minimize the amount of surgery and to perform it with extreme caution. The parts of the cochlea that can be accessed under these conditions include (1) those parts that can be seen through the round window at the very base of the cochlea (this is commonly referred to as the hook region of the cochlea) and (2) those parts that can have holes drilled into them without upsetting the rest of the cochlea too much. There are basically two of these, as illustrated in Figure 2.1B: one is partway around the first or most basal turn of the cochlea, and the other is partway around the last or most apical turn. Because the surgery that is necessary to expose even these parts of the cochlea is so invasive, all of the measurements that have been made to date have had to be made in deeply anesthetized animals. The anesthesia is thought to have little effect on the performance of the peripheral auditory system, however, and there are good reasons to believe that the best measurements made to date reflect almost “normal” mammalian hearing. Perhaps the most important of these reasons is that the best mechanical measurements have been supported by careful physiological controls. These show that it is possible to make mechanical measurements in cochleae that have undergone little or no change in their sensitivity to sound. And as we shall see later, the sensitivity of the cochlea is tightly linked to its amplification and compression of a sound.

24

N.P. Cooper

The second limitation on direct measurements of cochlear function is that they invariably involve making holes into the cochlea and usually require the placement of foreign bodies (e.g., radioactive sources or tiny lightreflecting mirrors) on the cochlear partition. The effects of making these holes and introducing these objects have been considered in both theoretical and practical terms. At present, the consensus view appears to be that holes made into the basal-turn scala tympani have little or no effect on cochlear function, whereas those made into the scala vestibuli (to access the apical turn of the cochlea, for example) can alter the tuning of the cochlear partition quite dramatically (e.g., see de Boer 1991; Ulfendahl et al. 1991; Cooper and Rhode 1996a). The foreign bodies are thought to have little influence on the results in most modern studies (e.g., see Sellick et al. 1983; Cooper 1999a), although they remain a cause for concern in some instances (e.g., see Khanna et al. 1998). 2.1.1 Input-Output Functions Perhaps the most straightforward way to illustrate the compression that occurs in the mechanics of the cochlea is to consider the variations in the amplitude of a mechanical response that occur when the intensity of a sound is varied. These variations are best illustrated in the form of inputoutput functions such as those shown in Figure 2.2. Each of the curves in Figure 2.2 shows the amplitude of a mechanical response to a sinusoidal stimulus (a tone) as a function of the intensity of that stimulus (measured in dB SPL; i.e., dB re: 20 mPa). The different curves in each panel of Figure 2.2 illustrate the responses to different frequencies of stimulation (relatively low-frequency data are plotted with open symbols and dashed lines and relatively high-frequency ones with solid symbols and solid lines), and the different panels illustrate responses observed in different species or in different regions of these species, cochleae. It should be evident immediately that there is a fair amount of variation in the observations made to date. Nonetheless, closer inspection of the data reveals many similarities across both species and region (or the “place” along the length of the partition). If we concentrate first on the data that were collected at the most sensitive or characteristic frequency (CF; as indicated by the solid circles and thick lines in Fig. 2.2) of each preparation, for example, we can see a fairly consistent pattern, varying only in quantitative detail from one location to another. All but one of the thick CF curves in Figure 2.2 begins by growing almost linearly (i.e., at a rate of 1 dB/dB) with increasing intensity and then gradually bends over to grow at a rate of less than 1 dB/dB. That is, the CF responses become compressively nonlinear once either the sound level or the response amplitude grows above some criterion level or breakpoint.The breakpoints for the CF data are marked with arrows in Figure 2.2A–E. In some cases, the CF curves show a tendency to bend over once again at even

2. Peripheral Compression

Displacement (nm)

0

2

10

20

40 60 80 Gerbil BM, CF=11500

4000 9000 11500 13000 15000

dB

/ dB

1

0

C 20

100 200 500 800 1000

2

10

40 60 80 Chinchilla TM, CF=500 dB

0

E

10

0

20

40 60 80 Intensity (dB SPL)

10

100

9000 13000 17000 18300 19000

B

dB

B/

1d

0

10

0 2

10

20

40 60 80 Chinchilla BM, CF=9750

4000 7000 9750 10500 11500

100

dB

B/

1d

0

10

D

100

B/

1d

2

100

10

0

Displacement (nm)

A

10

0

Displacement (nm)

B

/d

B 1d

Guinea-pig BM, CF=17000

Displacement (nm)

10

20000 28000 31000 33000 37000

0 Displacement (nm)

Displacement (nm)

Cat BM, CF=31000 2

25

20 50 75 300 500 800

2

10

40 60 80 Guinea-pig OC, CF=300

100

dB

B/

1d

0

F

10

0

20

40 60 80 Intensity (dB SPL)

100

Figure 2.2. Mechanical input-output functions in various species and regions of the cochlea. A: from the basilar membrane (BM) at the extreme base of the cat cochlea. A–F, keys specify stimulus frequencies in Hertz. Characteristic frequency (CF) data are shown by bold lines. Dashed lines indicate growth rates of 1 dB/dB, and arrows indicate compression thresholds or breakpoints. (Reprinted with permission from Cooper and Rhode 1992. Copyright © 1992 World Scientific Publishing Co. Pte. Ltd.) B: from the BM in the basal turn of the guinea pig cochlea. (Reprinted with permission from Nuttall and Dolan 1996. Copyright © 1996 Acoustical Society of America.) C: from the BM in the basal turn of the gerbil cochlea. (Data are from Cooper 2000.) D: from the BM in the basal turn of the chinchilla cochlea. (Reprinted with permission from Rhode and Recio 2000. Copyright © 2000 Acoustical Society of America.) E: from the tectorial membrane (TM) in the apical turn of the chinchilla cochlea. (Reprinted with permission from Cooper and Rhode 1997. Copyright © 1997 American Physiological Society.) F: from the organ of Corti (OC) in the apical turn of the guinea pig cochlea. (Data are from Cooper and Dong 2001.)

26

N.P. Cooper

higher intensities, but this time in the opposite direction to that observed at lower intensities (e.g., see Fig. 2.2C,E). That is, the slopes of some of the CF curves tend to increase back toward 1 dB/dB at the highest intensities shown (cf. Johnstone et al. 1986; Ruggero et al. 1997). At this point, it may be worth considering the relationship between compressive nonlinearity and compression as defined in Section 2.1. Not surprisingly, compression is a straightforward consequence of compressive nonlinearity: response growth rates of less than 1 dB/dB necessarily imply that a greater range of input signals are being processed to fit into a smaller amount of “output space” than would otherwise be possible. We can illustrate this by taking the CF input-output function in Figure 2.2C as an example. In this case, the input signals are quantified in terms of SPL, and the “output space” comes in the form of basilar membrane displacements. The CF responses in Figure 2.2C grow from around 3 nm at 30 dB SPL to just less than 30 nm at 90 dB SPL. Hence, a 60-dB range of input signals is being compressed into just less than a 20-dB (i.e., 10-fold) range of outputs. Both the degree (i.e., the shallow slope) and the wide dynamic range of this compression appear to be essential for normal sound processing in the cochlea. The absolute intensities at which the breakpoints between the linear and compressively nonlinear sections of the mechanical input-output functions occur vary considerably from species to species and from place to place in a given species (see arrows in Fig. 2.2). However, the basic pattern of the curves is the same in all cases. The only matter of real debate at present is how the compressive regions of the input-output functions relate to the overall, either physiologically or psychophysically relevant, intensity ranges. Arguably, the best mechanical data that are available to date imply that the lowest intensity breakpoints occur within 5 dB of the threshold of hearing (Nuttall and Dolan 1996; Ruggero et al. 1997; see Robles and Ruggero 2001). However, It should be noted that these implications are based on data from only two research laboratories, and that these laboratories make observations in just one region of the cochlea. The majority of observations made elsewhere suggest that the lowest intensity breakpoints occur around 20–25 dB above the threshold of hearing (this finding also proves to be more consistent with the implications that can be drawn from studies in the auditory nerve, as will be discussed later). There is a slightly better consensus regarding the highest intensity limit of the CF compression. Compression that extends to at least 100 dB SPL has been observed in at least five laboratories, and definitely appears to be a normal feature of hearing (Ruggero et al. 1997; Cooper 1998; de Boer and Nuttall 2000; Rhode and Recio 2000; Ren and Nuttall 2001; see Robles and Ruggero 2001 for a review). If we now look at the input-output functions for a given cochlear location as a function of frequency relative to the location’s CF, several other generalizations can be drawn from the data in Figure 2.2. In all but one case, for example, the curves relating to the below CF responses (the dashed lines

2. Peripheral Compression

27

with open symbols) exhibit less nonlinearity than is seen in the CF responses. That is, the below CF curves do not bend over either as much or at as low an intensity as the CF curves do. In fact, when the stimulus frequency falls sufficiently far below the CF of each site, most of the inputoutput functions become linear (i.e., grow at rates of 1 dB/dB) across the entire range of intensities observed. The only exceptions to this finding are seen in the data in Figure 2.2E,F, which come from the relatively littlestudied apical turn of the cochlea (these data are noteworthy in many respects, and are discussed further in Section 2.1.2). In contrast to the case for below CF responses, many of the input-output functions for tones presented just above the CF of each preparation show just as much, if not more, compression than is evident in the CF curves. In some cases, the above CF responses actually decrease in magnitude as the stimulus intensity increases, at least over a limited range of intensities (e.g., Fig. 2.2D,E). Some of the above CF input-output functions show a distinct transition from compressive to near linear or perhaps even slightly expansive behavior above around 80–90 dB SPL (cf. Fig. 2.2C–F, solid triangles). However, these transitions seem to depend on the exact frequencies tested, with responses at other frequencies exhibiting compression to the highest intensities tested (cf. Fig. 2.2B,C,E, solid diamonds). These characteristics resemble those expected of a system in which the responses originate from multiple pathways, as will be described in Section 2.2. A final response region, which is not illustrated in the data in Figure 2.2, is observed at frequencies well above the CF of each recording site. This region is known as the plateau region. The phases of the responses in this region show very little dependence on frequency. The amplitudes of the responses in the plateau region are very small, but they grow almost linearly (i.e., at a rate of 1 dB/dB) with increasing intensity. As discussed later, there is some debate over exactly what happens in the transition region between the frequencies illustrated in Figure 2.2 and those in the plateau region proper, but this debate is probably of little importance (see Section 2.1.2). 2.1.2 Quantitative Measurement of Compression The input-output functions in Section 2.1.1 illustrate the compression that occurs in cochlear mechanics in a very straightforward way, but they do not quantify this compression directly. To quantify the compression, the data have to be transformed into rate-of-growth functions. This could be achieved by simply differentiating the input-output functions. However, more information can be revealed to the human eye if the rates of growth are plotted as a function of stimulus frequency at each observation site. Examples of such plots are shown in Figures 2.3C,D and 2.4C,D, along with another series of readily interpretable plots (Figs. 2.3A,B and 2.4A,B) that show the tuning characteristics of the sites under study.

2

Displacement (nm)

10

N.P. Cooper

A

BF80

3

10

X X

1

X X X X

0

X

10

10

dB SPL 90 100 70 80 50 60 30 40 10 20

-1

10

-2

10

X

1

Displacement (nm)

28

CF 3

B

X

2

10

X X X

1

X

10

X X

0

10

CF

-1

10

10

0.1

Frequency (kHz)

CF

0.5

dB SPL 90-100 70-80 80-90 50-60 60-70 30-40 40-50 10-20 20-30

0

1

3

Frequency (kHz)

10

Growth rate (dB/dB)

Growth rate (dB/dB)

1

2

C

1

-0.5

0.3

Frequency (kHz)

2 1.5

BF80

X

1.5

D

CF

1 0.5 0 -0.5

0.1

0.3

1

Frequency (kHz)

Figure 2.3. Tuning and compression characteristics at 2 sites in the chinchilla cochlea. A and C: from the BM in the basal turn of the cochlea. (Adapted with permission from Rhode and Recio 2000. Copyright © 2000 Acoustical Society of America.) B and D: from the TM in the apical turn of the cochlea. (Adapted with permission from Cooper and Rhode 1997. Copyright © 1997 American Physiological Society.) A and B: tuning characteristics of each site shown by plotting the amplitudes of the vibrations evoked by various frequencies at intensities of between 10 and 100 dB SPL. ¥: Best frequencies (BFs) for each site as defined by the stimuli evoking the largest vibration velocity (not displacement) at each intensity. (e.g., BF80) C and D: quantification of the degree of compression that is evident across different intensity ranges (e.g., 50–60 dB SPL) at each site. Strong compression is indicated by low (i.e., closer to zero) growth rates, whereas linearity (i.e., the complete absence of compression) is indicated by growth rates of 1 dB/dB.

The data in Figures 2.3 and 2.4 were selected to illustrate the differences in both tuning and compression characteristics between sites in the two halves of the cochlea (sites near the base of the cochlea and sites near the apex). The data come from studies of the chinchilla and guinea pig cochlea. These are the most commonly used species for cochlear mechanics studies and are the only species where modern in vivo measurements have been made at both ends of the cochlea. In each case, there are notable differences between characteristics observed at the base and the apex of the

2. Peripheral Compression

29

2 3

A

10

BF90

X

Displacement (nm)

Displacement (nm)

10

1

10

X X X X X X

0

10

-1

10

10

-2

dB SPL 90 100 70 80 50 60 30 40 10 20 4

X X

CF 10

B

X 1

10

0

10

CF

-1

10

20

0.025

-0.5

Growth rate (dB/dB)

Growth rate (dB/dB)

0.4

1

2

C CF

1

0

0.1

Frequency (kHz)

2

0.5

BF90

X

2

10

Frequency (kHz)

1.5

X X

dB SPL 90-100 70-80 80-90 50-60 60-70 30-40 40-50 10-20 20-30 4

10

Frequency (kHz)

20

1.5

D CF

1 0.5 0 -0.5

0.025

0.1

0.4

1

Frequency (kHz)

Figure 2.4. Tuning and compression characteristics at 2 sites in the guinea pig cochlea. A and C: from the BM in the basal turn of the cochlea. (Adapted with permission from Cooper 1998. Copyright © 1998 The Physiological Society. [London]) B and D: from the organ of Corti in the apical turn of the cochlea. (Data are from Cooper and Dong 2001.)

cochlea; in general terms, the basal sites, which are tuned to high frequencies, reveal much stronger, much more extensive (i.e., covering a wider range of intensities), and much more sharply tuned compression than the apical sites. There is some concern about the validity of all the apical-turn data observed to date, however, so comparisons between the apical- and basal-turn characteristics must be made with caution. This is particularly so in the case of the guinea pig data in Figure 2.4B,D, where the signs of compression are very weak indeed. In the case of the apical-turn chinchilla data, the signs of compression are sufficiently strong and sufficiently widespread across frequency, that even if the amount of compression has been underestimated severely (by compromising the physiological condition of the cochlea in order to make the observations), the pattern of the compression can still be seen to differ from that in the more basal parts of the cochlea. The data in Figures 2.3C and 2.4C illustrate both the frequency and intensity dependence of the compression that occurs in the basal regions of the

30

N.P. Cooper

mammalian cochlea. The compression is indicated by a response growth rate of less than 1 dB/dB, and only becomes evident above a frequency that is around half an octave below the CF for the location under study (e.g., above approximately 6 kHz in Fig. 2.3C and approximately 12 kHz in Fig. 2.4C). The amount of compression generally increases with increasing frequency, albeit in an intensity-dependent manner, up to a point that is around one-quarter of an octave above the CF (e.g., approximately 12 kHz in Fig. 2. 3C and approximately 20 kHz in Fig. 2.4C). What happens beyond this point is a matter of some debate, but perhaps little importance. The reason for the debate is that different investigators find different results: some find that growth rates return quite rapidly to 1 dB/dB (e.g., Ruggero et al. 1997) as the response plateau region (cf. Section 2.1.1) is approached, whereas others find areas where the growth rates exceed 1 dB/dB (i.e., there is evidence of expansive nonlinearity) before they return to 1 dB/dB (e.g., Cooper and Rhode 1992; Rhode and Recio 2000). Whether this discrepancy is real, or whether it is simply a consequence of the different frequency resolutions that have been used in the different laboratories, remains to be seen. The reasons that any debate over “who is right” here can be viewed as unimportant in any case are that (1) the magnitudes of the mechanical displacements that are being considered are miniscule, and (2) the responses in question do not seem to be passed on to subsequent stages in the auditory periphery (cf. Cooper and Rhode 1996a; Narayan et al. 1998; see Robles and Ruggero 2001 for a review). The latter point raises an interesting question for those who study cochlear mechanics, but it is clearly unrelated to the issue of compression. The data in Figures 2.3D and 2.4D illustrate the frequency and intensity dependence of the compression that occurs in the apical regions of the mammalian cochlea. In the case of the chinchilla data in Figure 2.3D, some compression is evident across the entire frequency range of the measurements, extending from at least two octaves below the CF to at least one octave above the CF. At any given intensity of stimulation, the amount of compression varies only slightly with frequency below approximately 900 Hz. The amount of compression does vary systematically with intensity, however; in general, the compression is nonexistent (the growth rate approximates 1 dB/dB) at low intensities, maximal (around 0.5 dB/dB) at moderate intensities (e.g., between 40 and 60 dB SPL near the CF), and then moderate (between 0.5 and 1 dB/dB) at high intensities. It has been suggested that the amount of compression in the apical turn of the cochlea depends more on the level of the displacements of the partition rather than on the SPL per se (Rhode and Cooper 1996). However, this suggestion is not consistent with observations made in the more basal regions of the cochlea. In the case of the apical-turn guinea-pig data shown in Figure 2.4D, much less compression is evident at most frequencies within about one octave of the CF of the preparation, but there are clear hints of the patterns seen in

2. Peripheral Compression

31

both the apex of the chinchilla cochlea and the basal turns of many other cochleae. There is, for example, some compression even at the lowest frequencies tested (again over two octaves below the CF), just as there is in the apex of the chinchilla cochlea. However, little or no compression is evident at the CF of the guinea pig preparations (Cooper and Rhode 1995; Ulfendahl et al. 1996; Zinn et al. 2000). On the high-frequency side of the CF, there is evidence of strong compression (e.g., growth rates of just 0.2 dB/dB) over a very restricted frequency region, but this rapidly turns into a region where both compression and expansion (i.e., growth rates of greater than 1 dB/dB) can be seen at different intensity levels (Cooper and Rhode 1995). This is somewhat akin to the situation seen in the more basal turns of the cochlea. It is noteworthy that the growth rate of the responses to a few tones around one octave above the CF in Figures 2.3D and 2.4D is significantly greater than one. That is, there is clear evidence of expansive nonlinearity at sites near the apex of the cochlea (in both chinchillas and guinea pigs). The expansion in the apex cannot be “ignored” like that in the basal turns was because the actual amplitudes of the responses involved are not trivial. The expansion in the apical turns is almost always associated with inputoutput functions that contain sharp discontinuities, or notches, at a particular intensity (cf. Fig. 2.2E). This observation has been used by some investigators (e.g., Cooper and Rhode 1995, 1996a) to suggest that the expansion may be caused by the intensity-dependent interference of a compressively nonlinear process with a linear process. However, other investigators have taken their own observations of expansive nonlinearity as direct evidence for negative feedback in the mechanics of the cochlea (Zinn et al. 2000). These ideas are discussed at greater length in Section 2.2. 2.1.3 Dynamic Aspects of Compression All of the data considered so far have related to stationary properties of the peripheral auditory system. That is, they have been based on measurements of the performance of the system that were made without regard to the time course of the responses. As will become apparent later (cf. Fig. 2.11), the amount of compression that can be seen at subsequent stages in the periphery (e.g., in the auditory nerve) depends on the temporal as well as the spectral characteristics of a sound or a sound-evoked response. There is very little evidence to suggest that such time-dependent characteristics originate in the mechanics of the cochlea, however. In fact, there is strong evidence to the contrary. This evidence comes from several lines of experimentation. The first two lines of evidence come from looking at the time courses of the mechanical responses to single tones or click stimuli. The waveforms of the responses to individual tone bursts almost invariably show that the instantaneous responses of the system (measured while the tonal stimuli

32

N.P. Cooper

are rising or falling in amplitude, for example) undergo compression in just the same way that the steady-state responses do (e.g., see Rhode and Cooper 1996; Ruggero et al. 1997; Rhode and Recio 2000). The individual cycles of the response waveforms are not particularly distorted (Ruggero et al. 1997; Cooper 1998), but their envelopes seem to trace out exactly the same input-output functions as the time-averaged responses discussed in Section 2.1.1. The waveforms of the cochlear partition’s responses to click stimuli are somewhat more complicated than those of its responses to tones, but in essence they show the same characteristics. This is illustrated explicitly in Figure 2.5, which shows two series of click responses from the two halves of the chinchilla cochlea. The peak-to-peak (or, more accurately, peakto-trough) amplitudes of the individual cycles of the click responses have been plotted against the peak-equivalent intensity level of the click in Figure 2.5C,D, and compressive nonlinearity (i.e., growth rates of less than 1 dB/dB) is seen for every cycle. In the apical-turn data in Figure 2.5D, the degree of compression is small, but the compression is basically similar regardless of time. This is most probably a consequence of two factors: first, the frequency composition of each cycle in the apical-turn click responses is fairly similar, and second, as we saw in Figs. 2.2E and 2.3D, some compression occurs at almost all frequencies in the apex of the chinchilla cochlea. In the basal-turn data of Figure 2.5C, there are clear and systematic differences between the input-output functions for the individual cycles of the clickevoked responses, but these do not mean that the compression depends on time per se.The initial cycles of the basal-turn click responses show much less compression than the later cycles, but the differences turn out to be explicable entirely in terms of the frequency composition of the responses. The earliest cycles of the basal-turn click responses are driven by the lowest frequency components of the stimulus because these components travel along the cochlear partition much faster than the high-frequency components (there is a progressive shift in the frequency of the components that drive the click-evoked responses, from well below the CF of the preparation at the onset of the responses to somewhere very close to the CF after between 3 and 4 cycles; cf. Robles et al. 1976; de Boer and Nuttall 1997; Recio et al. 1998; Recio and Rhode 2000). The earliest cycles of the basal-turn click responses therefore grow at a relatively high rate against stimulus intensity (just as the responses to below CF tones do; cf. Sections 2.1.1 and 2.1.2), whereas the later cycles grow at a much lower rate, exhibiting a greater degree of compression over a much wider dynamic range. Further lines of evidence that mechanical compression does not depend on time come from the phenomenon of two-tone suppression. This is a phenomenon whereby the sensitivity of the responses to one tone (a probe tone, which is normally placed at the CF of the site under study) can be affected by the presence of a second tone (a suppressor tone, which may be either higher or lower in frequency than the probe tone). Findings from

2. Peripheral Compression

16

Displacement (nm)

±64

±158

96dB

±50

86dB

±16

76dB

±8

66dB

±4

umbo (116dB)

±16 0 Peak-to-Peak Displacement (nm)

3

4

95dB 85dB 75dB 65dB

±16

55dB

±5.0

45dB

±1.6

1

2 Time (ms)

3

dB

B/

100

1d 5 6 4 7 3

10 6 5 4 7 3 2 16

50

5 6 4 7 3 2 16 1

2 16 1

5 4 6 3 7 2 16 1

4 3 5 6 2 7 1 16

3 4 2 5 6 1 7

3 2 4 1 5 6 7 16

3 2 1 4 5 6 7 16

16

60 70 80 90 100 110 120 Intensity (p.e. dB SPL)

umbo (75dB)

±16

4

C

35dB

±0.5

46dB

±1

40

2

56dB

±2

1

1

±500

106dB

±32

1000

B

116dB

Displacement (nm)

1 2 3 4567

0 Peak-to-Peak Displacement (nm)

A ±128

33

1000

4

8 Time (ms)

16

D

1 2 3

dB

B/

100

1d

1 2 3 1 2 3

10

1 2 3 2 1 3

1

20

12

1 2 3

1 2 3 4

4

4

4

4

4

30

40 50 60 70 80 Intensity (p.e. dB SPL)

90 100

Figure 2.5. Compression characteristics revealed in the responses of the BM to click stimuli at 2 sites in the chinchilla cochlea. A and C: basal-turn data from Recio and Rhode (2000). (Reprinted with permission. Copyright © 2000 Acoustical Society of America.) B and D: apical-turn data from Cooper and Rhode (1996a). (Reprinted with permission. Copyright © 1996 Taylor & Francis Ltd [http://www.tandf.co.uk/journals].) A and B: BM waveforms in response to 10- and 100-ms condensation clicks, respectively, at the peak-equivalent (p.e.) intensities indicated. Bold lines: responses of the BM; thin lines: response patterns predicted by linear extrapolation from the highest level responses. The waveforms of the mechanical stimuli to the cochleae, as recorded from the malleus in the middle ear, are shown in the lowermost traces (umbo). C and D: input-output functions for the individual cycles of the click responses, with numbers indicating the different time periods of the responses illustrated in A and B, respectively. Dashed lines indicate growth rates of 1 dB/dB.

34

N.P. Cooper

A

two tone suppression

2nm

response to probe

response to suppressor

Displacement (nm)

B 1

10

20

B

/d

21dB 0.1

B 1d

31 36 41 51 61 71

0.01 20

40 60 Probe Level (dB SPL)

30ms

C

Probe = 18.8kHz Suppressor = 22.9kHz

80

Displacement (nm)

0

Probe = 800Hz Suppressor = 1kHz

100

dB

B/